RE: Next question...

["marcelo bagnulo" <marcelo@it.uc3m.es>]

Hi Tony,

> Folks,
>
> Have we reached consensus that we need to deal with multihoming
> policies?

I am sorry but i do not fully understand what do you mean by this.

That a site has to be able to define the exit router used for its packets?
That a site has to be able to set what ingress router will be used for the
response of the packets it sent before?
From an end-site point of view, something else?

From an ISP perspective, what considerations must be taken into account?

>  And do we agree that doing so at the per-host level
> doesn't scale?

I would say that the management of such solution doesn´t scale, i mean when
a site has more and more hosts, it becomes harder and harder to manage
policies, i guess.
So i guess that this type of solutions is not suitable for big sites with a
large number of hosts, but it would work fine for small sites. I mean is not
that this type of solution does not scale when it is widely used, only that
it is not suitable for very large sites, where i guess it will not be
adopted. So i would say that we can work on this type of solutions, but
other solutions are needed to address large site multi-homing.

Besides i think that this type of solutions presents very good scalability
when considering other aspects, such as storing multiple locators of a host.
I mean, storing at each host its multiple locators scales better than
storing this information in the routing system or in specific databases.

Regards, marcelo

>
> Thanks,
> Tony
>